A phytosociological survey of aquatic vegetation in the main freshwater lakes of Greece

Aims: This study aims to contribute to the knowledge of European freshwater lake ecosystems with updated and new information on aquatic plant communities, by conducting national-scale phytosociological research of freshwater lake vegetation in Greece. Moreover, it investigates the relationship between aquatic plant communities and lake environmental parameters, including eutrophication levels and hydro-morphological conditions. Study area: Lakes in Greece, SE Europe. Methods: 5,690 phytosociological relevés of aquatic vegetation were sampled in 18 freshwater lake ecosystems during 2013–2016. The relevés were subjected to hierarchical cluster and indicator species analyses in order to identify associations and communities of aquatic vegetation, as well as to describe their syntaxonomy. Multiple regression analysis was applied to investigate the relationship between vegetation syntaxa and environmental parameters of lakes, i.e. physico-chemical parameters and water level fluctuation. Results: Ninety-nine plant taxa belonging to 30 different families were recorded. Forty-six vegetation types were identified and described by their ecological characteristics, diagnostic taxa and syntaxonomical status. Thirteen vegetation types, the largest number belonging to the vegetation class Charetea, are considered to be new records for Greece. The distribution of the vegetation types recorded in the 18 freshwater lakes was found to depend on environmental parameters and levels of eutrophication. Conclusions: An updated aquatic vegetation inventory was produced for Greek lakes, and primary results showed that the presence/ absence of aquatic plant communities and the community composition in freshwater lakes can be utilized to assess the pressure of eutrophication on lake ecosystems. Taxonomic reference: Euro+Med (2006–). Abbreviations: MNT = Mean number of taxa; WFD = Water Framework Directive.


Introduction
Freshwater ecosystems are among the most threatened ecosystems around the world (Sala et al. 2000;Foley et al. 2005;Dudgeon et al. 2006). Overexploitation, water pollution, flow modification, destruction or degradation of habitats, and exotic species invasions are the five main drivers of biodiversity loss in freshwater ecosystems (Dudgeon et al. 2006). The European Union addressed the vulnerability of freshwater ecosystems with the adoption of the European Water Framework Directive (WFD, European Commission 2000). In this frame- work, the monitoring of aquatic plant communities was proposed as a key element in order to assess the ecological status of freshwater ecosystems, as macrophytes play a significant role in determining the structure and functions of lake ecosystems by influencing environmental conditions, nutrient cycling, and biotic assemblages and interactions (Carpenter and Lodge 1986;Jeppesen et al. 1997;Engelhardt and Ritchie 2001). As a result, most of the monitoring and assessment systems developed by European countries utilise rankings in the tolerance and sensitivity of macrophyte taxa to eutrophication (Kolada et al. 2014;Poikane et al. 2018). The monitoring of aquatic macrophytes in Greek freshwater ecosystems, in the context of the Greek National Water Monitoring Network (GNWMN) under the WFD, began in 2013 (Zervas et al. 2018).
The number of floristic and phytosociological investigations in freshwater ecosystems within Greece has increased during the past three to four decades (Sarika-Hatzinikolaou et al. 2003;Sarika et al. 2005). Also publications containing phytosociological data for lacustrine aquatic plant communities have accumulated over time, but remain scarce and not evenly distributed across the country: Gradstein and Smittenberg (1977: western Crete), Lavrentiades and Pavlidis (1985: Lake Mikri Prespa), Papastergiadou (1990: various lakes in Northern Greece), Bergmeier (2001: seasonal pools in the island of Gavdos), Sarika-Hatzinikolaou et al. (2003: seven lakes in Epirus), Grigoriadis et al. (2005: Agras wetland), Dimopoulos et al. (2005: Kalodiki marsh); Zotos (2006: Lakes Trichonida and Lysimachia), Fotiadis et al. (2008: Lake Chimaditida), and Pirini (2011: Lakes Vegoritida and Petres). These studies provide important information about aquatic vegetation in Greece, but the older ones do need to be revised and updated. Furthermore, research gaps remain in the country, i.e. a number of important lakes remain unsurveyed.
Taking into consideration all of the above information, the main objectives of this study are (i) to contribute to the knowledge of European freshwater lake ecosystems with new and updated country-wide information on the aquatic plant communities found in the main Greek freshwater lakes, and (ii) to investigate the relationship between the distribution patterns of macrophyte communities and environmental parameters indicating increased levels of eutrophication and altered hydro-morphological conditions.

Study area
The study covers 18 lakes (Table 1; Figure 1) selected for GNWMN monitoring of aquatic macrophytes (Mavromati et al. 2017;Zervas et al. 2018). While the studied lakes are scattered over the Greek mainland, most of them are clustered in the west and north-central part of the country, differing in altitude, size, water depth, and local climatic conditions within their catchment area (Table 1). Of the three transboundary lakes (Doirani, Megali Prespa, Mikri Prespa) only their Greek areas were studied.

Vegetation and environmental data
Each lake was surveyed once in 2013-2016 during the main growing season (May to September) (Table 1). In all lakes, the belt transect-mapping method was applied Table 1. Overview of the geographical, geometric and climatic characteristics of the studied lakes. Asterisks mark transboundary lakes, for which the characteristics refer to their part in Greece. Climatic characteristics have been collected by the European Climate Assessment & Dataset (Klein Tank et al. 2002). Average annual temperature and annual precipitation values have been calculated on the basis of available data during the period 1995-2005. Survey period and number of transects and relevés surveyed per lake is also given.  (Zervas et al. 2018), the most commonly used method for aquatic vegetation surveys in Europe, due to the fact that it provides abundance, frequency and depth distribution data for the different taxa found within the vegetation of a lake (Kolada et al. 2009). Ten to 20 transects per lake were established from the shoreline perpendicular to the maximum depth of plant growth. Sampling was conducted in relevés of 4 m 2 , evenly distributed along the belt transects following a gradient of increasing depth. Sampling was undertaken using a double-headed rake with a scaled handle or attached to a rope, a bathyscope, and a geo-bathymetric device. In this way, a total of 5,690 relevés were sampled, in which all angiosperms (helophytes, hydrophytes, amphiphytes and aquatic forms of land species), pteridophytes, bryophytes, charophytes and green filamentous macroalgae (e.g. Cladophora spp.) were recorded and determined to species or subspecies level (except filamentous macroalgae), and their abundance was estimated with the use of the semi-quantitative five-point DAFOR scale (Palmer et al. 1992). Vascular plant taxonomy follows Euro+Med (2006), while algae taxonomy follows Guiry and Guiry (2019). Chorological information was collected from Dimopoulos et al. (2013Dimopoulos et al. ( , 2016, Guiry and Guiry (2019), and Julve (1998). A number of environmental data (e.g. total phosphorus concentrations in the water column, Secchi depth, water electric conductivity, water level fluctuation measurements) were collected periodically from each lake in the context of GNWMN (for details see Zervas et al. 2018). These data were used to assess the relationships between the distribution patterns of aquatic syntaxa and eutrophication and hydro-morphological factors.

Statistical analysis
In order to define the vegetation types in the most objective manner possible, the relevés were subjected to a number of hierarchical cluster analyses. Extremely rare taxa, i.e. recorded in one to three out of 5690 plots, were excluded from the analyses in order to reduce "noise" in the data. DAFOR abundance classes were translated to their average percentage abundance values as follows: Dominant = 87.5%, Abundant = 50%, Frequent = 17.5%, Occasional = 5.5% and Rare = 0.5% (CEN 2007). Species abundances were chord distance-based transformed (Legendre and Galacher 2001). The transformed dataset was then subjected to cluster analysis with the use of flexible beta linkage method with b = -0.25 (Lance and Williams 1967) and Bray-Curtis dissimilarity (Bray and Curtis 1957). Elbow and Average Silhouette methods (Kaufman and Rousseeuw 1990), and NbClust statistic (Charrad et al. 2014) were used to assist in the determination of the optimal number of clusters for the dataset. Finally, diagnostic taxa were determined by indicator species analysis (Dufrene and Legendre 1997;De Cáceres et al. 2012), using the indicators function, in order to finalize the number of clusters corresponding to distinct vegetation types, and describe the best combination of indicator species for each vegetation type.
Due to the overall low number of common taxa among the resulting clusters, the hierarchic dendrogram that was produced was not able to successfully group all vegetation types into meaningful syntaxa, thus we proceeded with an additional cluster analysis. The synoptic table, which contained the clusters representing our dataset, was integrated into a dataset of clusters representing the types of Greek aquatic vegetation published in the past (bibliography in Suppl. material 1) and was processed again using the flexible beta linkage method and Bray-Curtis dissimilarity. The aim of including these vegetation types from the literature within our dataset was to support the present syntaxonomical decisions. The syntaxonomy of higher syntaxa (alliances, orders and classes) in the current study follows, with few exceptions, Mucina et al. (2016).
Depth distribution for each vegetation type was calculated and presented. The distribution of higher-rank syntaxa for each lake was also computed on the basis of the number of relevés per syntaxon in proportion to the total number of relevés in each lake. Calculations were summarized at the level of class for most of the vegetation types, except the ones belonging to the Potamogetonetea which were divided at the level of alliance, owing to the high variation in this class with different life forms. Finally, a multiple linear regression model was applied to assess the relation between aquatic vegetation patterns, as expressed by the abundance of higher-rank syntaxa, and environmental parameters in each lake. Pearson's correlation coefficient (R) and p-value (p) of the model were assessed.

Vegetation classification
Cluster analysis and subsequent tests resulted in 46 different vegetation types for interpretation (see Suppl. material 3 for Elbow, Average Silhouette and NbClust results, and Suppl. material 4 for produced dendrogram). Due to the survey methodology used, i.e. consecutive relevés distributed along a depth gradient at equal depth intervals, a number of the resulting vegetation types correspond to transitional ecotonal stands. These vegetation types were retained in the synoptic tables and are described in the text so as to present a more comprehensive picture of the spatial and ecological patterns of vegetation differentiation within the studied lakes. For syntaxonomic purposes, they may well be merged with an adjacent vegetation type. The diagnostic species for each vegetation type were selected from the results of the indicator species analysis as those combinations that reached a higher Indicator Value, while maintaining high prediction power and sensitivity (De Cáceres et al. 2012) (see Suppl. material 5 for all diagnostic taxa parameters). Diagnostic and accompanying species for each vegetation type are given in Tables 2-4. Short descriptions of the ecology (structure, water-depth preference etc.), the floristic composition and the distribution for each vegetation type are presented at the following paragraphs (see Suppl. material 6 for summary of vegetation types in all lakes). Syntaxonomic remarks that led to their final syntaxonomic assignment (Table 5) are also presented.

Class 1. Plantaginetea majoris
Syntaxon 1.1. Phyla nodiflora community (Code PhN, Table 2, Mean number of taxa MNT = 2.4) Appearance and habitat: Sparse temporarily submerged carpets, dominated by Phyla nodiflora, a perennial herb of prostrate growth, covering periodically flooded shores. Phyla nodiflora is a cosmopolitan pioneer herb that grows prolifically in floodplain wetlands with periodical flooding of short duration (Sharma and Singh 2013). Other aquatic macrophytes rapidly colonizing flooded areas, such as Myriophyllum spicatum and Vallisneria spiralis, can also be found in this community.
Diagnostic taxa (% constancy): Phyla nodiflora (100%  et al. 2009). In publications from the western Mediterranean basin (e.g. Brullo and Sciandrello 2006;Ninot et al. 2011) an association of Phyla nodiflora growing in littoral grassy plains together with Panicum repens (Lippio nodiflorae-Panicetum repentis O. Bolòs 1957) has been described, but our community differs as Panicum repens is absent. Our material is insufficient to provide a firm basis for describing a new association. We do not follow Mucina et al. (2016) who treat the perennial Phyla nodiflora as a diagnostic species of the class Isoëto-Nanojuncetea, defined as pioneer ephemeral vegetation in periodically flooded freshwater habitats. We assign the Phyla nodiflora community described here to the order Paspalo-Heleochloetalia and to the alliance Paspalo-Agrostion semiverticillati instead, which comprises Mediterranean-subtropical temporarily inundated, disturbed, perennial grass-herblands rich in stoloniferous plants of tropical and subtropical distribution. Table 2, MNT = 3.1) Appearance and habitat: Emerged and floating mats of Paspalum distichum colonizing exposed areas of wet ground that may be temporarily shallowly inundated. Paspalum distichum is a perennial grass, originating from tropical America, which is widely established in riparian habitats of the Mediterranean basin, often forming monotypic stands (Aguiar et al. 2005).
Distribution: Doirani, Lysimachia, Paralimni, Trichonida and Vegoritida. Table 2. Synoptic table of the identified associations and communities belonging to Classes Plantaginetea majoris, Phragmito-Magnocaricetea and Lemnetea. Taxa constancy in percentage and their average abundance class (r = 0-1%, + = 2-5%, 1 = 6-20%, 2 = 21-40, 3 = 41-60%, 4 = 61-80%, 5 = 81-100%) superscripted are shown. Companion taxa with less than 20% constancy are shown at the end of the    (José et al. 1988;Rivas-Martinez et al. 2001;Neto et al. 2009), we choose to assign our vegetation type as a variant of the first one, which is first in priority order if P. distichum dominance stands are treated as a single association. Zotos (2006) identified two communities with Paspalum distichum in his study of wet meadows around lakes Trichonida and Lysimachia, including one dominated by Paspalum distichum. All the above-mentioned associations and communities have been grouped in the alliance Paspalo-Agrostion semiverticillati and order Paspalo-Heleochloetalia. We do not follow Mucina et al. (2016) who grouped this order of perennial herb-grasslands in the annual-dominated class Bidentetea and we prefer the class of perennial plant communities on damp or temporarily flooded, often trampled, disturbed ground, Plantaginetea majoris, which Mucina et al. (2016) lumped together with the Molinio-Arrhenatheretea.

Class 2. Phragmito-Magnocaricetea
Syntaxon 2.1. Phragmitetum communis (Code PA, Table  2, MNT = 1.2) Appearance and habitat: Extensive and dense (>50% cover) reed beds of Phragmites australis, the most commonly noticed and recorded association in most lakes. They cover major parts of the littoral zone, reaching down to 6m depth.
Syntaxonomic remarks: This association, widespread across all bioclimatic zones of Eurasia, matches with what has been identified as Phragmitetum communis (australis) or Scirpo-Phragmitetum in numerous publications in Greece (Drosos et al. 1996;Sarika-Hatzinikolaou et al. 2003;Grigoriadis et al. 2005;Zotos 2006) and Europe (Preising et al. 1990;Šumberová et al. 2011a;Landucci et al. 2013;Kamberović et al. 2014;Jenačković 2017;Lastrucci et al. 2017 Appearance and habitat: Stands of Phragmites australis with floristic composition similar to the preceding cluster but with lower Phragmites cover (<50%). They are found at the edges of dense reed beds, down to 6m depth, where the Phragmitetum communis progressively gives way to, or is interconnected with, aquatic communities such as Cladophoretum glomeratae, Najadetum marinae, Lemnetum minoris, Ceratophylletum demersi, Potamogetono pectinati-Myriophylletum spicati etc. Due to their sparse cover, other riparian and aquatic plants of the above-mentioned or other plant communities colonize the open areas among and beneath the reeds.
Syntaxonomic remarks: This cluster falls within the range of variation of the Phragmitetum communis. Syntaxon 2.3. Scirpetum lacustris (Code SL, Table 2, MNT = 5.6) Appearance and habitat: Dense stands of club-rush Schoenoplectus lacustris (>25% cover) and low presence of other helophytes (Phragmites, Sparganium and Typha spp.). In lacustrine ecosystems, it often forms a zone in mostly shallow waters down to 1m deep, sensitive to wave action, between the open water and the dense reed-bed areas dominated by other species, like Phragmites australis.

Class 3. Lemnetea
Syntaxon 3.1. Lemnetum minoris (Code LM, Table 2, MNT = 5.8) Appearance and habitat: Mats of the free-floating duckweed Lemna minor (>50% cover), accompanied by less abundant lemnids, such as Spirodela polyrhiza, Azolla filiculoides and other Lemna spp., can be found in the littoral zone of still and relatively nutrient-rich freshwater bodies, in very shallow waters 0-1m deep, in spots protected against wave action.
Syntaxonomic remarks: Matches the descriptions of this widespread association from Greece (Lavrentiades and Pavlidis 1985;Papastergiadou 1990;Zotos 2006) and elsewhere in Europe (Goldyn et al. 2005;Kłosowski and Jabłońska 2009;Šumberová 2011b;Felzines 2012 Appearance and habitat: Open to fully closed submerged carpets of the free-floating carnivorous bladderworts Utricularia vulgaris or Utricularia australis (>25% cover), with other taxa found in low numbers. As the bladderworts cannot be identified with certainty if not in flower, both species are likely to be included. Frequently present at the surface of the water occur Hydrocharis morsus-ranae and lemnids, like Lemna minor, Lemna gibba, Spirodela polyrhiza etc., while Ceratophyllum demersum may occur in lower strata of the water column. Vegetation of free-floating bladderworts can be found in very shallow, down to 1m deep, mesotrophic to eutrophic waters protected against wave action.
Syntaxonomic remarks: Matches the descriptions of this widespread association from Greece (Sarika-Hatzinikolaou et al. 2003;Pirini 2011, with Utricularia vulgaris and Chara vulgaris) and elsewhere in Europe (Šumberová 2011b;Felzines 2012;Džigurski et al. 2016;Cvijanović et al. 2018). Table 2, MNT = 1.5) Appearance and habitat: Extensive (>50% cover) carpets of Ceratophyllum demersum, a free-floating aquatic macrophyte in variable habitat conditions. Due to its ability to grow well in turbid water, under poor light conditions, it spreads rapidly and may cover the whole water column, possibly limiting the growth of other hydrophytes. While it thrives mostly in shallow waters, it may colonize the full depth range of aquatic macrophytes (in Greece 0-13m).
Syntaxonomic remarks: Matches the descriptions in European publications (Goldyn et al. 2005;Šumberová 2011b;Felzines 2012;Lastrucci et al. 2014Lastrucci et al. , 2015Džigurski et al. 2016;Cvijanović et al. 2018). In Greece, Papastergiadou (1990) and Dimopoulos et al. (2005) identified this association with similar floristic composition, while Sarika-Hatzinikolaou et al. (2003) described a more variable and perhaps composite association, with higher constancies of other Lemnetea and Potamogetonetea diagnostic taxa (Lemna minor, Spirodela polyrhiza, Hydrocharis morsus-ranae, Myriophyllum spicatum and Potamogeton crispus). Gradstein and Smittenberg (1977)  Syntaxonomic remarks: These complex stands may be assigned to any of the two associations depending on species' prevalence. Appearance and habitat: Dense stands (mostly >50% cover) of the water-milfoil Myriophyllum spicatum, a submerged macrophyte with a broad ecological range, common even in disturbed sites. It roots at the lake bottom and reaches the water surface to emerge its inflorescence. These stands colonize waters down to 6m deep, provided water transparency is sufficiently high (chiefly mesotrophic conditions).
Distribution: Amvrakia, Feneos and Paralimni. Syntaxonomic remarks: Matches the descriptions of this widespread but infrequent association (Melendo et al. 2003;Šumberová 2011a;Lastrucci et al. 2014;Džigurski et al. 2016;Cvijanović et al. 2018), which in Greece, so far only Papastergiadou (1990, as Ranunculetum fluitantis but with similar floristic composition) described in slow-flowing waters. Appearance and habitat: Dense stands (>25% cover) of the submerged pondweed Potamogeton compressus accompanied at lower abundance by taxa such as Vallisneria spiralis, Stuckenia pectinata and Najas marina. Its shallow root system is vulnerable to wave action, thus Potamogeton compressus forms limited stands in shallow (down to 2m deep) water near lake shorelines.
Distribution: Kastoria. Syntaxonomic remarks: Only a few publications described this association from Eurasia (Kuzmichev et al. 2008;Borsukevych 2013;Chepinoga et al. 2013), which is rare and/or declining in Europe (Birkinshaw et al. 2013). There are no previous records of this association from Greece. Syntaxon 4.(1.)13. Potamogetonetum trichoidis (Code PT, Table 3, MNT = 6.1) Appearance and habitat: Dense stands (>25% cover) of the submerged narrow-leaved pondweed Potamogeton trichoides, accompanied at lower abundance by taxa such as Myriophyllum spicatum, Ceratophyllum demersum and Lemna minor. Being quite variable, this vegetation type was found in meso-eutrophic waters down to 4m deep, where Potamogeton trichoides leaves spaces for a mix of other elodeid and lemnid aquatic macrophytes as well as helophytes.

Class 4. Potamogetonetea: Alliance 2. Nymphaeion albae
Syntaxon 4.(2.)17. Trapetum natantis (Code TN, Table 3, MNT = 4.3) Appearance and habitat: Open to closed (>25% cover) floating mats of the annual water caltrop Trapa natans, most often accompanied by Ceratophyllum demersum which tolerates poor light conditions. Nymphaeids such as Trapa natans are macrophytes that root at the bottom of still freshwater bodies, but most of their biomass, in particular most of the leaves, is floating on the water surface. Trapa occurs in waters down to 3m deep, limiting light levels for other submerged macrophytes underneath.
Distribution: Kastoria and Megali Prespa. Syntaxonomic remarks: The Trapetum natantis has been described in Greece, (Lavrentiades and Pavlidis 1985;Papastergiadou 1990) and Europe (Šumberová 2011a;Džigurski et al. 2016;Cvijanović et al. 2018). Appearance and habitat: Open to closed (>25% cover) floating vegetation mats of the water lily Nymphaea alba, most often accompanied by Ceratophyllum demersum which is undemanding in terms of light. Like other nymphaeids, Nymphaea alba is bottom-rooted and forms dense floating leaf mats, occurring in waters down to 4m deep.
Syntaxon 4.(2.)21. Ludwigia peploides community (Code LP, Table 3, MNT = 2.0) Appearance and habitat: Open to closed (>25% cover) mats of Ludwigia peploides subsp. montevidensis, an amphibious perennial macrophyte forming creeping mats on the wet mud and flooded shores of freshwater bodies or floating mats on the muddy surface of the riparian zone. The floating mats, often found within the gaps of Phragmites australis reedbeds, reach down to 2m deep, leaving no room for other aquatic macrophytes.
Diagnostic taxa (% constancy): Ludwigia peploides ssp. montevidensis (100%). Distribution: Lysimachia. Syntaxonomic remarks: Ludwigia peploides subsp. montevidensis, native to South America, is locally naturalized in South Europe, SW Asia and other continents where it is often invasive (Dutartre 1986;Zotos et al. 2006). In South America the association Polygono-Ludwigietum peploidis has been described (Padovani et al. 1993;Hauenstein et al. 2002), where Ludwigia peploides is often (but not always) accompanied by Persicaria hydropiperoides which does not occur in Europe. We did not find Ludwigia peploides relevés from Europe other than those published by Zotos (2006) and Zotos et al. (2006), together with Paspalum distichum or dominated by Phragmites australis. We found Ludwigia peploides as the dominant species associated with Phragmites. Taking into consideration the ecological similarities between Ludwigia peploides and Ludwigia grandiflora ), a diagnostic taxon of the Nymphaeion, we assign with some reservations the Ludwigia peploides community to that alliance.
Distribution: Feneos. Syntaxonomic remarks: Matches the descriptions of this association from publications in Europe Iakushenko and Borysova 2012;Azzella et al. 2013). In Greece, to our knowledge, no distinct Chara globularis community has been hitherto identified.
Distribution: Vegoritida. Syntaxonomic remarks: Matches the descriptions of this association in Europe (Hrivnák 2002;Iakushenko and Borysova 2012;Täuscher and van de Weyer 2015). In Greece, a community dominated by Nitella mucronata has not yet been identified.  (Golub et al. 1991;Landucci et al. 2011;Csiky et al. 2014). In Greece, no community dominated by Nitella hyalina has been identified yet.

Class 7. Stigeoclonietea tenuis
Syntaxon 7.1. Cladophoretum glomeratae, lake-substratum variant (Code ClGL, Table 4, MNT = 2.3) Appearance and habitat: Open to closed (>25% cover) submerged carpets of the filamentous macroalgae Clado- phora glomerata, found in stagnant eutrophic lowland waters. It is a quite light-demanding taxon which is often entangled with other macrophytes (subsequent cluster), or attached to the rocky substrate. These relevés, with a low cover of other aquatic macrophytes, were found in waters down to 5m deep.

Relation of phytosociological units to environmental parameters
Water depth is widely known to be an important environmental parameter which affects the distribution of aquatic plants, by regulating prevailing light conditions, temperature, water chemistry, wave action and substrate granulometry (Spence and Chrystal 1970;Chambers and Kaiff 1985;Middelboe and Markager 1997). Each macrophyte species has its own water depth tolerance limits, which depend on its morphological and physiological characteristics. However, due to the competition for space, light and nutrients from other macrophyte species they are not free to colonize the water volume that falls within their tolerance limits (McCreary 1991;Gopal and Goel 1993;Gross 2003). These mechanisms produce distinct zonation patterns in aquatic vegetation along water depth gradients (Spence 1982;Shipley et al. 1991). Figure 2 summarizes the depth distribution of the 46 described vegetation types, as recorded in the lakes that were surveyed in the current study. Among the helophytic vegetation types (Plantaginetea majoris; Phragmito-Magnocaricetea) the Phyla nodiflora community, and the Paspalo distichi-Agrostietum verticillatae, Scirpetum lacustris, and Typhetum angustifoliae were recorded colonizing the littoral zones to a depth of 1.5m. The Typhetum domingensis, Typhetum latifoliae, and Butometum umbellati were able to reach a bit deeper down to a depth of 2m, while the Phragmitetum communis which dominates the littoral zone of Greek lakes, quite often reach-   es down to a depth of 4m. Freely floating macrophytes (Lemnetea) and anchored floating macrophytes (Nymphaeion albae) are also restricted to shallow waters down to 1m and 3m deep respectively, with the exception of the Ceratophylletum demersi which can be found commonly down to 6m deep. Submerged hydrophytes (Potamogetonion; Charetea intermediae) predominantly colonize the deeper part of the euphotic zone of lacustrine littoral areas, between the zone colonized by emergent vegetation and the aphotic zone. Therefore, the majority of vegetation types belonging to Potamogetonion or Charetea intermediae are usually located in a depth zone starting at 1-2m and reaching 4-6m deep (in Greek waters), depending on the variability of light penetration and the specific lake physico-chemical characteristics. In cases where the euphotic zone reaches more than 6-8m deep, the Potamogetonetum pectinati, Nitellopsidetum obtusae, and Charetum vulgaris are the most commonly found vegetation types. An equally important environmental parameter to water depth, that influences the distribution of aquatic plants, is prevailing light conditions. Light penetration in lacustrine ecosystems is highly dependent upon their water quality status (Phillips et al. 1978;Canfield et al. 1985;Middelboe and Markager 1997). Nutrient loading and eutrophication lead to the growth of phytoplankton, epiphytes and filamentous algae, which leads to increased shading and light attenuation. As a result, macrophyte dominance is reduced due to their biomass decline, plant cover reduction and loss of species richness (Phillips et al. 1978;Sand-Jensen 2000). Figure 3 and Table 6 summarize the relationships we found between the distribution and abundance of higher-rank syntaxa for each lake and the prevailing physico-chemical and hydrological conditions. Positive and significant correlations were found between the distribution of Phragmito-Magnocaricetea and Nymphaeion albae with total phosphorus concentrations, while Potamogetonion was negatively correlated. In addition, positive and significant correlations were found between Charetea intermediae and Platyhypnidio-Fontinalietea antipyreticae with Secchi depth transparency, while Phragmito-Magnocaricetea was negatively correlated. Only Potamogetonion was positively correlated with electrical conductivity. No syntaxon was correlated significantly with water level fluctuation. Multiple linear regression analysis produced the best solution for the above-mentioned environmental parameters (TP, SD and EC) using the combination of distribution values for five syntaxa: Phragmito-Magnocaricetea, Potamogetonion, Nymphaeion albae, Charetea intermediae, and Platyhypnidio-Fontinalietea. The distribution patterns of these five higher-rank syntaxa appear to act as good indicators of lake eutrophication. Raised total phosphorus concentrations in lake water and lowered water transparency led to the dominance of Phragmito-Magnocaricetea, and Nymphaeion albae syntaxa in aquatic vegetation. The expansion of Potamogetonion, Charetea intermediae, and Platyhypnidio-Fontinalietea syntaxa in aquatic vegetation is associated with lower total phosphorus concentrations and higher values of water transparency.
These results are of relevance for WFD assessment purposes and are similar to those presented in Poikane et al. (2018) that reviewed national macrophyte-based approaches for assessing ecological status according to the WFD. Poikane et al. (2018) reported that a marked decline in submerged vegetation, especially Charophyta (characterizing 'good' status according to WFD), and an increase in abundance of floating and emerged plants (characterizing 'less than good' status) were the most significant changes along the ecological status gradient. Similar results have also been reported from other areas within Europe, where the indicator value of different groups of taxa belonging to these syntaxa were tested against eutrophication levels in the context of WFD assessement systems (e.g. Penning et al. 2008aPenning et al. , 2008bSøndergaard et al. 2010;Kolada 2016). Table 6. Overview of the relationships between the abundance of higher-rank syntaxa (classes to alliances) for each lake within the current study and its environmental variables. Pearson's correlation coefficient (R) and the p-value of significance are given for each linear regression. Significant relationships (p < 0.05) are marked in bold. The two final rows of the table contain part of the results of the multiple linear regression analysis with the involvement of more than one higher-rank syntaxa (one with all the higher-rank syntaxa and one with those giving the best solution for all the environmental parameters). PLA: Plantaginetea majoris; PHR: Phragmito-Magnocaricetea; LEM: Lemnetea; POTA: Potamogetonion; POTB Nymphaeion albae; FON: Platyhypnidio-Fontinalietea antipyreticae; CHA: Charetea intermediae; STI: Stigeoclonietea tenuis; TP: Annual mean total phosphorus (μg/L); SD: Secchi depth transparency in meters; EC: Electrical conductivity (μS/cm); WLF: Annual water level fluctuation in meters. Syntaxa

Conclusions
The current study is a national-scale phytosociological survey of freshwater lake vegetation, based on the most recent data available (years 2013-2016). Forty-six vegetation types were identified and interpreted for eighteen major Greek freshwater lakes. Among these vegetation types, the following are new records for Greece: Phyla nodiflora community, Butometum umbellati, Potamogetonetum denso-nodosi, Potamogetonetum compressi, Najadetum minoris, Fontinaletum antipyreticae, Charetum globularis, Magno-Charetum hispidae, Nitellopsidetum obtusae, Charetum asperae, Nitelletum mucronatae, Nitelletum hyalinae, Cladophoretum glomeratae. A primary analysis on the distribution of higher-rank syntaxa of the 46 vegetation types showed that the majority of these types are significantly affected by physico-chemical parameters indicative of higher levels of eutrophication. Aquatic plant communities could be utilized in eutrophication indices to broaden the assessment of the ecological status of freshwater lakes. Additional research on this topic is needed.

Data availability
The data that support the findings of this study were used under license from The Goulandris Natural History Museum, Greek Biotope/Wetland Centre (EKBY). They are available from the lead author upon reasonable request and with permission of The Goulandris Natural History Museum, Greek Biotope/Wetland Centre (EKBY).
Gross EM (2003) Allelopathy of aquatic autotrophs. Critical Reviews in